Quanta: Open-Source Python Library Unifies Quantized Vectors, BM25 and Graph RAG
_reachsumit · x · 2026-09-17
An arXiv paper presents Quanta, a self-contained open-source Python library that unifies 4-bit quantized dense vector search, BM25 full-text retrieval, and knowledge-graph traversal behind one retrieval API — replacing the typical multi-system RAG stack (ANN index + search engine + graph DB + document store).
Key design choices:
- Weighted reciprocal rank fusion instead of normalizing heterogeneous scores onto a shared range, which the author argues is ill-posed because such normalizations are query-dependent.
- Graph as candidate expander, not relevance scorer: traversal widens the candidate pool; newly admitted documents are re-scored by the dense indexes under an identifier allowlist, so structure decides what is considered while content evidence decides ranking.
More from coding & agent
- Don't use an agent where a sequence works: the biggest AI-building mistake, per Hugo Bowne — hugobowne · 2026-09-17
- Plasma AI launches Radio: a shared chat room for Claude, Codex and Grok agents — rohanpaul_ai · 2026-09-17
- UIUC Launches Center for Orchestrated Agentic Biology to Study Brain Aging with Multi-Agent AI — dotey · 2026-09-17
- Dev uses LLMs to extract structured data from corporate financing documents — mayfer · 2026-09-17
- Open-source Claude skill builds MCP servers end-to-end via mcpfy SDK — ZealousidealTax42 · 2026-09-17
- First Jev test: 90% agreement with verified Gemini Flash workflow, not yet a drop-in replacement — CompleteSkeptic · 2026-09-17